Question 1: Capital Bikeshare Data
- Load in the data containing trip information from the Capital Bikeshare program. Also load in the station information. Rename variables that have spaces in the names trip data, station data
Note: These data were originally from http://data.codefordc.org/group/transportation.
stations <- read_csv(file = "https://dcgerard.github.io/stat_412_612/data/capital_stations.csv")
##
## ── Column specification ────────────────────────────────────────────────────────
## cols(
## id = col_double(),
## name = col_character(),
## terminalName = col_double(),
## lastCommWithServer = col_double(),
## lat = col_double(),
## long = col_double(),
## installed = col_logical(),
## installDate = col_double(),
## removalDate = col_double(),
## temporary = col_logical(),
## public = col_logical(),
## capacity = col_double()
## )
head(stations)
trips2016 <- read_csv(file = "https://dcgerard.github.io/stat_412_612/data/capital_trips_2016.csv")
##
## ── Column specification ────────────────────────────────────────────────────────
## cols(
## `Duration (ms)` = col_double(),
## `Start date` = col_character(),
## `End date` = col_character(),
## `Start station number` = col_double(),
## `Start station` = col_character(),
## `End station number` = col_double(),
## `End station` = col_character(),
## `Bike number` = col_character(),
## `Member Type` = col_character()
## )
trips2016 %>%
rename(Duration = "Duration (ms)", Start_date = "Start date",
End_date = "End date", Start_station_number = "Start station number",
Start_station = "Start station", End_station_number = "End station number",
End_station = "End station", Bike_number = "Bike number",
Member_Type = "Member Type") ->
trips2016
head(trips2016)
- Parse the date-time information from the trip data. Recall the times are recorded in the America/New_York time zone, not the UTC time zone. Specify that in your parser.
trips2016 %>%
mutate(Start_date = mdy_hm(Start_date, tz = "America/New_York"),
End_date = mdy_hm(End_date, tz = "America/New_York")) ->
trips2016
head(trips2016)
- Calculate the average number of trips for each weekday (Sunday, Monday, Tuesday . . . ) given the day has trips. There are several days with no trips.
• Save the resulting days of week and corresponding average number of trips as a data frame called sumdf and print it out.
library(lubridate)
trips2016 %>%
separate(Start_date, into = c("Start_d", "Start_t"), sep = " ") %>%
group_by(Start_d)%>%
summarize(trips = n()) %>%
mutate(wday = wday(Start_d, label = TRUE)) %>%
group_by(wday) %>%
summarize(mean_num_trips = mean(trips)) -> sumdf
print(sumdf)
## # A tibble: 7 x 2
## wday mean_num_trips
## <ord> <dbl>
## 1 Sun 5111.
## 2 Mon 6057.
## 3 Tue 6617.
## 4 Wed 6846.
## 5 Thu 7309.
## 6 Fri 6358.
## 7 Sat 6027
# Another Method: using yday()
# mutate(Start_date = yday(Start_date)) %>%
# count(Start_date)
- Reproduce this plot in R:
sumdf %>%
mutate(wday = ordered(wday, levels= c("Sat", "Fri", "Thu",
"Wed", "Tue", "Mon", "Sun"))) ->
sumdf1
ggplot(sumdf1, aes(x = mean_num_trips, y = wday)) +
geom_point() +
theme_bw()+
xlab("Mean Number of Trips") +
ylab("Weekday")
- In a stunning show of contempt, the IEEE Computer Society decided to add a new weekday called “Fooday” with abbreviation “Foo”. Fooday was decided to be the first day of the week (ahead of Sunday).
On the first Fooday ever, people used Capital Bikeshare in record numbers, yielding 15567 trips. Add Fooday as the first level to the wday variable in sumdf and add its average number of trips (now 15567 since there has only been one Fooday so far).
Hint: Create a new data frame that contains the Fooday trips and use bind_rows().
Fooday <- tribble(~wday, ~mean_num_trips,
##----/------------
"Foo", 15567)
# Fooday <- data.frame(wday = "Foo", mean_num_trips = 15567)
wday_vec <- c("Foo","Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat" )
Fooday %>%
mutate(wday = factor(wday, levels = wday_vec)) -> Fooday2
sumdf %>%
mutate(wday = factor(wday, levels = wday_vec, ordered = FALSE )) -> sumdf2
bind_rows(Fooday2, sumdf2) -> Fooday_trips
print(Fooday_trips)
## # A tibble: 8 x 2
## wday mean_num_trips
## <fct> <dbl>
## 1 Foo 15567
## 2 Sun 5111.
## 3 Mon 6057.
## 4 Tue 6617.
## 5 Wed 6846.
## 6 Thu 7309.
## 7 Fri 6358.
## 8 Sat 6027
- In another stunning show of contempt, the IEEE Computer Society decided to change the abbreviations from three letters to two letters. Change the levels of wday so that each day uses only two-letter abbreviations. Your final data frame should look like this:
Fooday_trips %>%
mutate(wday = str_sub(wday,1,2),
wday = parse_factor(wday))->
ABBR_Fooday_trips
print(ABBR_Fooday_trips)
## # A tibble: 8 x 2
## wday mean_num_trips
## <fct> <dbl>
## 1 Fo 15567
## 2 Su 5111.
## 3 Mo 6057.
## 4 Tu 6617.
## 5 We 6846.
## 6 Th 7309.
## 7 Fr 6358.
## 8 Sa 6027
- In the stations data frame, it seems that installDate is populated by the number of milliseconds since January 1, 1970, 00:00:00 (in the America/New_York time zone). Parse this into a date-time and make a histogram of the install dates. It should look something like this:
# 1000 milliseconds = 1 seconds
stations <- read_csv(file = "./data/capital_stations.csv")
##
## ── Column specification ────────────────────────────────────────────────────────
## cols(
## id = col_double(),
## name = col_character(),
## terminalName = col_double(),
## lastCommWithServer = col_double(),
## lat = col_double(),
## long = col_double(),
## installed = col_logical(),
## installDate = col_double(),
## removalDate = col_double(),
## temporary = col_logical(),
## public = col_logical(),
## capacity = col_double()
## )
sample_n(stations,6)
stations %>%
mutate(oldtime = ymd_hms("1970-01-01 00:00:00", tz = "America/New_York"),
installDate = (oldtime + dmilliseconds(installDate))) %>%
ggplot(aes(x = installDate)) +
geom_histogram() +
theme_bw() +
xlab("Install Date") +
ylab("Count")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
